HASH_SORT2(head,cmpfcn,cmpcls) about uthash HOT 4 CLOSED

You can do this today with HASH_SORT, or at least that's the intention:

struct my_order *order;
#define compare(a,b) cmp_func(order, a, b)
HASH_SORT(hash, compare);
#undef compare

So unless there's a case that that approach can't cover, I'm inclined to close this issue.

Btw, stylistically I'd recommend passing the cookie cls as your third argument, not the first, just because that's what the cookie-taking comparison function for qsort_r does.

from uthash.

I'm not sure if a macro will help me because I need to export uthash as library to use it on other language that doesn't provide macros support, but I'm going to test it anyway and back with feed back ...

from uthash.

I noticed you mentioned a Python binding elsewhere. I'm curious to see how that's going to work anyway. But given that uthash is pretty much 100% macros, one more little macro (compare(a,b)) in your own code shouldn't be a problem. I think either all the macros are going to give you trouble, or none are.
Let me know how it goes.

from uthash.

Arthur, your suggestion solved my problem and now I have a sort callback with closure support. :-)

BF_EXTERN bool bf_hs_sort(struct bf_hs **hsl, bf_hs_sort_cb cmp_cb, void *cmp_cls) __nonnull((1, 2));

...

bool bf_hs_sort(struct bf_hs **hsl, bf_hs_sort_cb cmp_cb, void *cmp_cls) {
    if (NULL == hsl || NULL == cmp_cb)
        return false;
#define CMP(a, b) cmp_cb(cmp_cls, a, b)
    HASH_SORT(*hsl, CMP);
#undef CMP
    return true;
}

Now I'm going to check the other issue about top/next.

Thank you very much dude! :-)

from uthash.